The present study highlights the impact of the living environment on the gut microbiome of healthy dogs. We found that the gut microbiome structure and function (α- and β-diversity and the differential abundances) were different between the household (Pet) and the non-household dogs (Kennel and UIUC), as well as between the two non-household cohorts even when adjusting for identical diet (fresh)/diet type (kibble). Dogs in the Kennel cohort received a single uniform kibble diet while the UIUC cohort received one of three kibble offerings. Dogs in the Pet cohort consumed various kibble brands and formulas.
Most published canine diet and gut microbiome studies have been conducted in a controlled facility environment 18,27–29, and their findings are often assumed to be transferable to household pet dogs. The results herein emphasize that a direct translation of findings from non-household to household dogs needs to be done with caution. The Pet cohort had a wider range of β-diversity along the PCoA axis 1 than the two non-household cohorts for both diet types (Fig. 3). Compared to dogs in a controlled environment, household pet dogs are usually exposed to a wide variety of microbial sources such as soil 30, water 31, other animal species, and dogs and humans from the same and different households 32,33, as well as dietary supplements, medications, treats, table scraps, and environmental compounds that may have an impact on the gut microbiome profile such as chemicals in household products, pesticides, and heavy metals 34. To our surprise, α-diversity was significantly higher in Kennel than Pet in both diets, even when living in a more uniform living environment in the Kennel. This could be that Kennel dogs were research animals and thus regularly exposed to a wider array of test diets and supplements. A wide range of Shannon H has been previously observed in household dogs across different U.S. geographical regions, ranging from lower than 1.0 to higher than 3.0 22, although this cohort of dogs included various breeds and the 16S rRNA sequencing was performed. For the two non-household cohorts, the difference in their exercise regime, feeding cadence (once vs twice a day) 35, cohabitation and socialization (group- or pair-housed vs individually housed) 36,37, and sanitization protocol 38, may have contributed to their dissimilar gut microbiomes. Our findings in these two non-household groups strongly emphasize the value of repeating research studies at different facilities to assure consistent findings before reaching a conclusion.
A unique component of this study was our ability to control for diet in the different cohorts, and therefore the impact of diet on the microbiome was completely (fresh) or partially (kibble) eliminated. Studies in other animal species comparing captive and wild populations examine the microbial impact of changing their living environments, but diet variability is often part of the research question 19,20. Likewise, significant differences in the gut microbiome were also observed in healthy humans living in different geographical regions within the same country 39–48, as well as before and after immigrating to a new country 49, but diet was, again, usually considered as a part of the changing living environment and was not controlled for. When both diet and genetics were controlled for in an experiment using genetically identical adult C57BL/6 mice, significant divergence in both the composition and metabolism of gut microbiota was observed when they were housed in separate controlled units within a single facility 50. Another unique aspect of this study was the use of solely purebred beagle dogs, in an attempt to minimize the effect of breed on the gut microbiome 51–53. However, the genetic variety of pet beagles and those commonly used in research may differ. Only one pet beagle in this study was acquired from a university or a laboratory. The original population of Beagles at Kennel were bred for research purposes and acquired from outside laboratories, and subsequent generations were bred in-house. That being said, a study in Hungary found no significant difference in the genetic diversity and no evidence that pet and kenneled beagles originated from different genetic pools 54, however, all beagles utilized in our study resided in North America.
The gut microbiome in this study was characterized using whole-genome shotgun metagenomic sequencing, a technique that is relatively new in canine microbiome research and provides higher accuracy and resolution than the more widely-used method of 16S rRNA sequencing 55. However, even at the phylum level, the most abundant phylum in both the fresh and kibble diets was still inconsistent among the three cohorts. The characterization of the gut microbiome in healthy pet dogs fed kibble has been previously performed with whole-genome shotgun metagenomic sequencing and Firmicutes is often found to be the first or second most abundant phylum 56–59, similar to what we observed in the Kennel cohort and the Pet cohort who were fed a wide variety of kibbles. In contrast, Actinobacteriota was the most abundant phylum in UIUC for a kibble diet. When healthy dogs were fed a fresh diet, Proteobacteria was a highly abundant phylum in the canine fecal microbiome in several studies 60–63, as was observed in our Pet and UIUC cohorts. However most of these prior studies were also conducted in U.S. household dogs or at UIUC. When beagles at Kennel were fed a fresh diet, Proteobacteria accounted for only 0.48% (median). This observation calls attention to the need for more diverse samples using this increasingly popular diet type 64,65. Despite the sullied reputation of Proteobacteria for including a number of well-known opportunistic pathogens, a 2018 review by Moon et al. critically discussed its diversity and importance in the gut of healthy dogs and cats 66.
The abundances of many species in the phyla Firmicutes_A, Firmicutes, Bacteroidota, Proteobacteria, and Actinobacteriota were mutually identified to be differentially abundant by both LEfSe and generalized linear models among the three cohorts. To ensure the robustness and accuracy of our results, we followed the recommendation by Nearing et al. and reported overlapping results from multiple differential abundance methods 67. For instance, probiotic species in the genus Bifidobacterium (in the phylum Actinobacteria) or in the family Lactobacillaceae (in the phylum Firmicutes) were found to be less abundant in Pet than Kennel and UIUC in both fresh and kibble diets. However, it is worth noting that all dogs included in this study were generally healthy. In humans, the gut microbiome of healthy individuals may share certain characteristics, such as being more resistant and resilient to disruption 68, but microbial communities are highly individualized, and a healthy microbiome cannot be defined by a single idealized composition 68. Findings from our study also support the complexity of defining a healthy core microbiome 69, and if such a microbiome can even be defined in dogs with diverse diets, genetic backgrounds, and geographic locations.
Intriguingly, the functional capacity of the gut microbiome was also different among the three cohorts. Taxonomic composition can be different between compared groups while their genetic composition or functional capacity may still be similar. This property, known as functional redundancy, has been hypothesized to underlie the stability and resilience of the gut microbiome in response to perturbations 70. Since the metagenomic functionality of the gut microbiome is closely related to changes in clinical symptoms and disease progression of the host 71, it further underscores the importance of understanding and evaluating environmental factors when studying health and the microbiome. Dogs with idiopathic inflammatory bowel disease 72, acute diarrhea 73, exocrine pancreatic insufficiency 74, and overweightness and obesity 9,11, were observed to possess dissimilar sets of metagenomic functionality compared to their healthy counterparts. Our findings carry implications for the well-being of dogs since diet and supplements can modify their gut microbiome functionality and may improve health 75–77. However, given that the dogs included in this study were absent of any diseases and the impact of these interventions may vary depending on the health status and the living environment, further research is needed to fully understand these nuances.
Additional limitations in the findings should be mentioned. Since this study is a cross-sectional study, there are always unknown variables that may confound the relationship between the cohort and the gut microbiome profile. For example, all beagles in the UIUC cohort were females while both sexes were included in the other two cohorts. It was suggested that sex and spayed/neutered status may have an impact on the canine gut microbiome 78. The feeding protocol and period were also not standardized across the cohorts, but each test diet was fed for at least seven days before stool collection which has been previously shown to be sufficient for the fecal microbiome to stabilize after diet change 79. Future studies may incorporate a longitudinal design to track the trajectory of the microbiome in different living environments.
In conclusion, the present study emphasizes the need for future investigations examining the effect of diet on canine gut microbiome and health to be inclusive of various living environments. Given the majority of diet and microbiome research is currently done in dogs living in controlled environments and then applied to wider populations, this study suggests microbiome research should also be conducted within household pets for greatest applicability.